388 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY BROWN, J. W. 1984. Host records for Para try tone melane (Edwards) (Hesperiidae). J. Lepid. Soc. 38:138. BURNS, J. M. 1992. Genitalic recasting of Poanes and Para try tone (Hesperiidae). J. Lepid. Soc. 46:1-23. COMSTOCK, J. A. & c. M. DAMMERS. 1931. Notes on the life history of Poanes melane Edw. (Lepid.). Bull. South. Calif. Acad. Sci. 30:20-22. EMMEL, T. C. & J. F. EMMEL. 1973. Butterflies of southern California. Nat. Hist. Mus. Los Angeles Co., Sci. Ser. 26:1-148. HEPPNER, J. B. 1972. The distribution of Paratrytone melane and its spread into San Diego County (Hesperiidae). J. Res. Lepid. 10:287-300. LANGSTON, R. 1980. Season Summary. News Lepid. Soc. 1980(2):14. SCOTT, J. A. 1986. The butterflies of North America. Stanford Univ. Press, Stanford, CA. 583 pp. RAYMOND V. BARBEHENN, Department of Entomological SCiences, University of California, Berkeley, California 94720, USA; current address: Department of Biology, University of Michigan, Ann Arbor, Michigan 48109-1048, USA. Received for publication 10 December 1993; revised and accepted 25 June 1994. Journal of the Lepidopterists' Society 48(4), 1994,388-393 COMMENTS ON THE NATURE AND ORIGINS OF MIGRATIONS OF LEPIDOPTERA TO BERMUDA Additional key words: evolution, migratory cues. "The Lepidoptera of Bermuda" (Ferguson, Hilburn & Wright 1991), which I recently reviewed for this Journal (Gaskin 1993), includes an interesting essay in the Appendix which, for brevity, I refer to below as Ferguson (1991). In this, he discussed a number of problems and paradoxes concerning the nature and origins of the long-distance mi­ grations of Lepidoptera to Bermuda, stimulating the additional thoughts and comments presented in this note. I will start by summarizing the general findings of Ferguson et al. (1991): Bermuda is an oceanic archipelago, dominated by one large island, that caps a seamount with no earlier geological connections to the North American mainland. This archipelago supports a disparate, super-saturated lepidopterous fauna, assembled by over-water dispersal at various times during the last five hundred thousand years, from southeastern North America and the northeastern Caribbean. A small number of endemic species are rec­ ognized, derived from ancestors in these two regions. The remaining members of the Bermudian fauna are morphologically and probably genetically, similar to their source populations. The biomass, diversity, and composition of the Bermudian fauna must have undergone dramatic increases and reductions during the Quaternary concomitant with several radical fluctuations in the surface area of the archipelago. It was not the primary purpose of Ferguson and his co-workers to develop hypotheses concerning the role that interactions of environmental stimuli, physiology, and behavior might have played in the evolution of long distance migrations of Lepidoptera to Bermuda. Certainly, the phenomenon appears to pose problems for the evolutionary geneticist. If there is no demonstrable return flight, then there seems to be no way for the trait to be fed back into the source population, and no cumulative selection from one generation to the next (Ehrlich 1984). Some of the Lepidoptera which migrate frequently to Bermuda have near-global ranges, and are found also on other oceanic islands. Others have close relatives in widely distant regions with similar migratory behavior. Ferguson estimates VOLUME 48, NUMBER 4 389 that the migrants to Bermuda make nonstop flights of more than 1000 km, often using lower altitude elements of the jet stream. A number of species from the southeastern United States found on Bermuda also make well-documented annual migrations into the northern states and Canada. Overwintering is unlikely in such cases because these species do not exhibit diapause. Ferguson speculates that many of these make return southern flights within continental North America that have not yet been detected. Presently the best known two-way migration within North America of course, is that of the monarch butterfly (Danaus plexippus L.; Nymphalidae). The movements have been copiously documented and the basic elements of the migration have been known for many years (Urquhart and Urquhart 1978). Northward spring flights in North America by species of Vanessa are also well known (Scott 1986). In 1992, for example, I counted about 125 V. cardui and V. virginiensis moving through one small part of Guelph, Ontario, Canada within about 4-5 hours on 26 April, pausing to nectar at clumps of spring dandelions on small patches of waste ground. Weaker southward return flights by V. cardui have been observed during late summer and fall in various parts of the United States (Scott 1986). There are also recent reports of consistent southerly autumn movements of V. atalanta in the United Kingdom (Riley & Riley 1992). Naturally, these observations may reflect only the prevailing wind direction at the time. Without resorting to mass alar tagging, it is exceedingly difficult to prove conclusively that observed movements of butterflies are in fact part of a large-scale migration rather than periodic local activity. It is, alas, quite unlikely that the kind of observer effort generated for the monarch butterfly pro­ grams could be duplicated for large-scale tagging of cutworm moths. Recent advances in methods to identify and trace specific combinations of isotopes through food chains may give us an additional indirect way of detecting long-distance migrations. Systematic sampling of mtDNA and nDNA in given species at different localities also may provide insight into population structure and movements. Unfortu­ nately, absolute methods for satisfactorily distinguishing sampled entities at the population level have yet to be perfected for animal species, despite great success in individual DNA "fingerprinting. " There are other ways of viewing the problems posed by Ferguson, however. It is rarely clear what proportion of any insect population migrates. We probably have more data on the migration of D. plexippus than for any other butterfly or moth, but some aspects are still unclear. Although huge numbers of monarch butterflies migrate from the southern wintering grounds into the northern United States and Canada in normal years, some still can be found even in southwestern Texas in the summer months. It seems unlikely that the migratory instinct is based on an "all or nothing" response in an entire population (Ehrlich 1984). Obviously long-distance return migrations, especially when two genera­ tions are involved as in the case of the monarch, demand the transfer of quite sophisticated cue recognition and navigational programming (Baker 1984, Douglas 1986). We have no clear idea of the time scale necessary for such developments. Ferguson's second puzzle is that the monarch butterflies of Bermuda are now residential, and no longer migrate. The migratory impulse presumably can be over-ridden in some way by internal systems that can recognize specific environmental cues or stimuli of immediate importance to the organism, attention to which will benefit the individual in the short-term and increase its chances of passing on its genetic heritage to offspring. Ignoring for a moment the fact that Bermudian monarchs are on an oceanic island, the question can be seen as part of the more general one "why don't colonial butterflies migrate?" Presumably, the majority tends to remain in the area through repeated exposure to local cues, such as the presence of specific microclimatic conditions or food plants that are absent for some distance around the colony. Statistically, relatively weak fliers carrying out short-distance search movements (random walks) will usually end up back in the locality where the cues are strongest or most repetitive. It is quite normal, however, for some individuals to disperse up to tens of kilometers from a colony. This happens in the case of the European large copper, Lycaena dispar (f. batavus), for example (Pullin et al. 1993). If such wanderers fail to find their way back and the colony is effectively isolated by unfavorable habitat, then this behavior could potentially threaten the viability of the colony. In other seasons the same behavior might enhance its survival by reducing 390 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY pressure on limited food resources. Under normal circumstances, however, in the absence of significant vagility (capacity for dispersal at least to the nearest suitable habitat site), local colonies of a few hundred individuals are certain to be extirpated in the long run by environmental events, often simple density-independent occurrences such as flood, drought, or unseasonable hard frost at some critical point in the life cycle. Stochastic events probably playa greater role in colony extinctions than inbreeding. Dempster (1989) argued that vagility, as a basic attribute, is vital for survival even for butterflies well­ adapted to highly specific habitats, because unless isolated colonies are "topped up" at intervals by surplus production from adjacent neighboring units experiencing "good years," they ultimately will disappear. Colonial butterflies in many parts of North America and Western and Eastern Europe have undergone dramatic declines in recent decades (Kudrna 1986). The endangered or threatened species are usually those with rather low vagility, often staying close to localized food plants. This was not such a problem before the advent of drastic modifications to northern temperate ecosystems